develooper Front page | perl.perl5.porters | Postings from February 2003

Re: Bug in perliol.pod

Thread Previous | Thread Next
February 10, 2003 17:00
Re: Bug in perliol.pod
Message ID:
BG = Benjamin Goldberg, NIS = Nick Ing-Simmons, HMB = H.Merijn Brand,
NC = Nicholas Clark between them write:

BG> In the description of struct _PerlIO_funcs, the Pushed member is:
BG>    IV   (*Pushed)(pTHX_ PerlIO *f, const char *mode, SV *arg, PerlIO_funcs *tab);
BG> But when it is described further on (in =item Pushed), it is:
BG>    IV   (*Pushed)(pTHX_ PerlIO *f, const char *mode, SV *arg);
BG> Checking the source (perlio.c), the former (with 'tab') is correct,
BG> and the latter is wrong.  As I am not particularly eloquent, who can
BG> suggest a good description of the 'tab' argument for the pod?

NIS> PerlIO_funcs is the structure describing the layer. (tab is short for table
NIS> here). Once upon a time PerlIO_funcs only had functions hence its name.
NIS> Now it is probably too late to s/PerlIO_funcs/PerlIO_layer/

HMB> My chainsaw is still warm. Should I ? :):)

NIS> Ask Jarkko/Hugo and PerlIO aware module owners e.g. Nick C.

NC> Last I knew it was me (2), Stas Bekman (with an Apache IO layer) and
NC> Dan Kogai (Encode) that were the only 3 owners of modules that do layers.

NIS> And I took :encoding back ;-)

NC> I don't give an $expletive - there were enough changes to track during the
NC> 5.7 series that doing this now isn't a problem for me, provided there is
NC> some way to work out what's going on, such as changing
NC> PERL_API_SUBVERSION from 0 to 1, or keeping that constant and saying that it
NC> comes in at PERL_VERSION 8, PERL_SUBVERSION 1 or later. 5.9.0 is allowed to
NC> change without warning - anyone not on p5p who is using it is silly.

NIS> So lets see what pumpkings think.

I'm happy with it, but it probably makes sense only if Jarkko would be
happy to take it into 5.8.1.


Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About